AeroLoop · Engineering note

AeroLoop compares landing response without adding flights

The repaired approach study records 12/18 fixed-gain and 15/18 scheduled-gain mission passes. A four-flight inspection lab keeps failed acceptance gates visible.

Source: https://louijiecompo.com/updates/aeroloop-landing-response/

Mission completion and acceptance are separate
Controller	Missions	Outage pairs	Recovery windows
Fixed gains	12/18	7/9	14/18
Scheduled gains	15/18	8/9	15/18

FIG 01 / Original acceptance gates across nine seeds. Derived from the retained approach study. Simulation only; accepted: false. PR #30 analyzes these flights and adds no measurements.

Inspect the pinned approach study ↗

Download the retained numeric report

PR #28 repairs clean-checkout test imports without changing recorded flights. PR #30 adds synchronized inspection of all 36 retained PhysX flights: fixed and scheduled gains, each with intact and two-second-outage feedback, across nine seeds. Its integrated CPU check passed.

The approach study records missions at 12/18 versus 15/18, outage pairs at 7/9 versus 8/9 and recovery windows at 14/18 versus 15/18. All-cohort accepted: false. The candidate remains opt-in, with failed support-position and recovery gates retained.

The response lab adds common-phase windows, control and work analysis, a nine-seed table and CSV/JSON exports. It adds zero flights and changes no acceptance criterion. Outage contrasts are descriptive, work residual is not measured friction, and full-rate bundles remain local upstream. The portfolio presents the public numeric report, not a new run or the response-lab playback.

Explore the topic

Autonomous systems under test

Follow the evidence

Explore AeroLoop

Read source evidence ↗ (opens in a new tab)

Explore the topic

Follow the evidence

Related reading